Neurosteroid secretion in panic disorder

Psychiatry Research 118 (2003) 107–116

Neurosteroid secretion in panic disorder Francesca Brambillaa,*, Giovanni Biggiob, Maria Giuseppina Pisub, Laura Bellodia, Giampaolo Pernaa, Vesna Bogdanovich-Djukica, Robert H. Purdyc, Mariangela Serrab a

Dipartimento di Scienze Neuropsichiche, Istituto Scientifico Ospedale S. Raffaele, Universita’ Vita e Salute, Milan, Italy b Dipartimento di Biologia Sperimentale ‘B. Loddo’, Cittadella Universitaria, Monserrato, Cagliari, Italy c Department of Psychiatry, Veterans Affairs Medical Center, University of California, San Diego, CA, USA Received 18 July 2002; received in revised form 4 March 2003; accepted 6 March 2003

Abstract Evidence that neurosteroids have anxiolytic effects in animal models of anxiety has suggested that alterations of neurosteroid secretion might be implicated in the pathogenetic mechanisms of anxiety disorders in humans. In 25 female patients with panic disorder (PD) and 11 healthy female controls, we measured plasma concentrations of progesterone (PROG), pregnenolone (PREG), allopregnanolone (3a,5a-tetrahydroprogesterones3a,5a-THPROG), dehydroepiandrosterone (DHEA) and tetrahydrodeoxycorticosterone (3a,5a-THDOC) during a drug-free month and during the following month of paroxetine therapy. The neurosteroids were measured during the early follicular phase, the mid-luteal phase and the premenstrual phase of both months (days 7, 22 and 27 from the beginning of the cycle). Significantly higher levels in patients than controls were found in PROG during the mid-luteal phase of both months, PREG in the premenstrual phase in the drug-free month, 3a,5a-THPROG during the follicular phase of the drugfree month and during the premenstrual phase of the therapy month, and 3a,5a-THDOC during the premenstrual phases of both months. DHEA levels did not differ in patients and controls. These results suggest that neurosteroids in PD are hypersecreted, possibly as an attempt to counteract the anxiogenic underlying hyperactivity of the hypothalamo-pituitary-adrenal axis and to improve a reduced GABAA receptor sensitivity. 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Panic disorder; Pregnenolone; Progesterone; Allopregnanolone; Dehydroepiandrosterone; Tetrahydrodeoxycorticosterone

1. Introduction In animal models neurosteroids influence cognitive function and behavior associated with emotional state probably through allosteric modulation of the GABAA ybenzodiazepine receptor function in the central nervous system (Crawley et al., *Corresponding author. Centro di Psiconeuroendocrinologia, Piazza Grandi 3, 20129 Milan, Italy. Fax: q39-2-70122889. E-mail address: [email protected] (F. Brambilla).

1986; Flood et al., 1988, 1990, 1992; Rupprecht and Holsboer, 2001). In fact, brain and plasma concentrations of pregnenolone (PREG), progesterone (PROG), dehydroepiandrosterone (DHEA), deoxycorticosterone, their metabolites and their sulphate derivatives have been observed to modulate in animals behavioral and biochemical responses to acute and chronic stress, anxiety, depression, aggressivity, convulsivity, anesthesia, sleep, memory, pain and feeding behavior (for

0165-1781/03/$ - see front matter 䊚 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0165-1781(03)00077-5

108

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

review, see Biggio and Purdy, 2001). These observations have suggested that fluctuations of neurosteroids might be involved in the development, course and prognosis of some mental disorders. This has been demonstrated in major depressive disorders (MDD) (George et al., 1994; Wolkowitz et al., 1997, 1999; Heuser et al., 1998; Romeo et al., 1998; Rupprecht and Holsboer, 2001; Uzunova et al., 1998; Strohle et al., 1999, 2000; Young et al., 2002), in premenstrual dysphoria (Wang et al., 1996, 2001; Rapkin et al., 1997; Bicikova et al., 1998; Daly et al., 2000; Rasgon et al., 2001), in anorexia and bulimia nervosa (Monteleone et al., 2001), in Alzheimer’s disease and in vascular dementia (Bernardi et al., 2000), where increased, decreased or dysregulated secretion of the main neurosteroids and their metabolites has been observed, the impairments correlating with some of the psychopathological aspects of the mental disorders. There are few data on neurosteroid secretion in simple anxiety and in anxiety disorders. Postmenopausal women with symptomatic mild depressive symptomatology and marked anxiety have been reported to have concentrations of DHEA and of its derivative 5a-androstane3a,17b-diol that are significantly lower than those of asymptomatic subjects (Barbaccia et al., 2000), with a significant negative correlation between neurosteroid levels and severity of anxiety. In patients with generalized anxiety disorder, serum levels of PREG sulphate have been found to be significantly lower than normal, with a trend toward low concentrations for DHEA and allopregnanolone (3a,5a-THPROG) in men and a trend toward high concentrations of DHEA in women (Semeniuk et al., 2001). Reduced plasma concentrations of PREG sulphate with normal levels of 3a,5a-THPROG and DHEA have been found in men suffering from generalized social ´ phobia (Heydari and Le Melledo, 2002), while in social anxiety DHEA sulphate plasma levels have been reported to be normal (Spivak et al., 2000). In posttraumatic stress disorder, DHEA sulphate plasma levels have been reported to be normal (Spivak et al., 2000). In a small group of men and women with panic disorder (PD), Strohle et al. (2002) observed significantly higher than normal

concentrations of 3a,5a-THPROG and 3a,5btetrahydroprogesterone, and lower than normal concentrations of 3b,5a-tetrahydroprogesterone, with normal concentrations of PROG, 5a-dihydro-PROG and 5b-dihydro-PROG. The values of the neurosteroids did not change after 24 weeks of paroxetine therapy. Tait et al. (2002), in a study in a small group of male PD patients with panic attacks artificially induced by pentagastrin, a cholecystokinin-B agonist, observed a significant release of DHEA, but only a trend toward a lower release of allopregnanolone in patients than in healthy controls. Clinical evidence and data from controlled studies suggest that gonadal steroid hormones, possibly including neurosteroid precursors, might influence the symptomatological expression of PD, by their ability to modulate respiration sensitivity (Klein, 1993). Most of the studies, but not all (Stein et al., 1989; Cook et al., 1990), have reported that PD is exacerbated in the premenstrual period (Breier et al., 1986; Sandberg et al., 1986; Cameron et al., 1988) and in the postpartum period, with lactation and pregnancy apparently being protective (Klein, 1964; George et al., 1987; Metz et al., 1988; Cowley and Roy-Byrne, 1989; Villeponteaux et al., 1992; Klein et al., 1995). In addition, some case reports showed that oral contraceptives and estrogen replacement therapy were able to trigger panic attacks (Price and Heil, 1988; Deci et al., 1992). All these data suggest that physiological fluctuations of some of the sexual hormones might modulate the occurrence, frequency and severity of panic attacks, and of the anxious symptomatology that is present in the syndrome. In a previous study, we observed that 35% CO2 inhalation-provoked panic attacks in panic patients were more frequent, and anxiety significantly more severe, in the early follicular than in the mid-luteal phase of the cycle (Perna et al., 1995). This suggested that fluctuations of sexual steroids in the menstrual cycle and their eventual secretory alterations might modulate the reactivity to CO2. Moreover, they might influence the suffocation alarm threshold, which is supposed to be impaired in patients with PD, possibly being a cofactor for the development of the attacks. Therefore, we decided to study in a preliminary inves-

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

tigation whether or not the secretion of PROG, PREG, 3a,5a-THPROG, DHEA and 3a,5aTHDOC, five neurosteroids that in animal models have been revealed to be highly effective anxiolytics, might be abnormal in PD during the early follicular, mid-luteal and premenstrual phases, in drug-free conditions and during treatments with the selective serotonin reuptake inhibitor (SSRI) paroxetine. Moreover, we also evaluated whether or not these hormonal alterations could be related to the anxiety and to the panic symptomatology, and if they disappeared during successful psychopharmacological therapy. In fact, recent clinical studies have shown that SSRIs modulate neurosteroid synthesis, and that clinically effective treatments with fluoxetine increase 3a,5a-THPROG concentrations in the cerebrospinal fluid of depressed patients, in whom the neurosteroid levels are lower than normal (Romeo et al., 1998; Uzunova et al., 1998). 2. Methods and materials Twenty-five female outpatients with PD and agoraphobia without other Axis I psychiatric disorders entered the study. They were 19–45 years old (mean age"S.D.s33.3"7.6), with histories of the disorder of 1–28 years (mean"S.D.s 6.3"6.5). In the 15 days before starting the study, the weekly frequency of panic attacks retrospectively measured by the panic-associated symptom scale (PASS) was 2.6"2.8. The patients were recruited over 1 year at the Anxiety Disorder Clinical and Research Unit of the Department of Neuropsychiatric Sciences, Istituto Scientifico Ospedale S. Raffaele, Milan, Italy. Consensus diagnoses according to DSM-IV criteria were obtained by two senior psychiatrists who assessed patients independently by clinical interviews and by the Mini International Neuropsychiatric Interview-Plus (MINI) (Sheehan et al., 1994). Before our study, all the patients had been treated with benzodiazepines, which were stopped at least 25 days before our investigation began. No one had been given SSRIs in the 6 months preceding the study. Eleven physically healthy female volunteers matched for age, recruited from the hospital staff, formed the control group. They were also inter-

109

viewed with the MINI to exclude present or lifetime Axis I disorders. Patients and controls were free of present medical illnesses, as evidenced by their medical histories, physical examination and laboratory assessments, and had histories of regular menstrual cyclicity (mean duration of menstrual cycles was 29"0.9 days). Exclusion criteria for patients and controls were organic diseases, immunopathies, allergopathies, inflammatory disorders, endocrinopathies, obesity or recent weight loss, alcohol or drug abuse, use of contraceptive drugs, organic brain disorders, cerebral trauma, and seizure disorders. Patients and controls gave informed consent to participate to the study. The day before each blood drawing for neurosteroid assays, measures of psychopathological symptomatology were obtained with the PASS (Argyle et al., 1991), the fear questionnaire total scores (FQtot; Marks and Mathews, 1979) and the state-trait anxiety inventory (STAI; Spielberger et al., 1970). The biological and psychological investigations in both patients and controls started at the beginning of the menstrual cycle for the drugfree month and for the following month under therapy. Paroxetine administration started on the first day of the menstrual period of the therapy month. During the treatment month, 20 mg of paroxetine were administered orally, once a day. Edetic acid (EDTA)-anticoagulated blood was drawn to measure basal concentrations of PREG, PROG, 3a,5a-THPROG, DHEA, and 3a,5aTHDOC, three times during the drug-free month of observation (A) and three times during the following month, in which the patients were treated with the SSRI paroxetine (B). During each month, neurosteroids were measured during the early follicular phase (A1, B1), in the mid-luteal phase (A2, B2) and in the premenstrual phase (A3, B3), that is, on days 7, 22 and 27 from the beginning of the cycle. In controls, they were measured only during 1 month, in A1, A2 and A3 phases. During the drug-free month patients were given a placebo, administered orally once a day, in a single-blind design. Controls did not receive paroxetine or placebo. For the blood collection for neurosteroid assays, patients arrived at the day hospital of our institute

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

110

Table 1 Psychopathological aspects of panic patients Rating scales

A1

A2

A3

B1

B2

B3

PASS FQtot STAI

6.7"3.8 24.8"14.6 47.2"10.0

5.6"2.9 15.0"9.0*** 44.2"10 g

6.6"3.5 14.8"11.4 44.7"11.3

6.5"3.3 7.8"10.3*** 46.3"11.8

2.2"2.0§ 15.4"10.2 35.3"7.7**

2.1"2.2l*** 7.4"7.0* 36.4"8.4 d*

Means"S.D.; PASS, Panic associated Symptom Scale; FQtot, fear questionnaire total scores; STAI, State-Trait Anxiety Inventory; A1, drug-free cycle, day 7; A2, drug-free cycle, day 22; A3, drug-free cycle, day 27; B1, therapy cycle, day 7; B2, therapy cycle, day 22; B3, therapy cycle, day 27; B vs. A, *P-0.001; **P-0.007; ***P-0.0004; §P-0.0009.

at 08.00 h, after 12 h of fasting, and rested in a supine position for 30 min before blood was drawn. At 08.30 h a cannula was inserted into an antecubital vein, kept patent by saline infusion, and at 09.00 h EDTA-anticoagulated blood was drawn, immediately centrifuged and the plasma frozen at y20 8C until assayed. For the neurosteroid assays 1 ml of plasma was diluted with 2 ml of water and then extracted three times with 3 ml ethyl acetate. The recovery (90%) of steroids through the extraction procedures was monitored by adding a trace amount of tritiated w3Hx cortisol (6000–8000 cpm 52 Ciymmol, New England Nuclear). The neurosteroids were quantified by radioimmunoassay as previously described (Serra et al., 1999), with specific antibodies to PROG, PREG, DHEA, 3a,5a-THPROG, 3a,5aTHDOC (UCN, Costa Mesa, CA). Antibodies to 3a,5a-THDOC and 3a,5a-THPROG were generated in rabbits and sheep, respectively, and were provided by Dr R.H. Purdy (Scripps Research Institute, La Jolla, CA). The specificity of the antibody to 3a,5a-THPROG, which showed no cross-reactivity with other steroids including THP3b and THP-5b, had been previously characterized (Purdy et al., 1991a). The limit of detection of the radioimmunoassays, expressed as minimal amount of steroids distinguishable from the zero sample, was 0.01 hg. Intra- and inter-assay coefficients of variation ranged between 5 and 7% and between 9 and 11%, respectively. The significance of the data was evaluated statistically by parametric Wilcoxon analysis to compare psychological aspects of patients and of controls. To compare neurosteroid concentrations in patients and controls, we used analysis of variance (ANOVA) for repeated measures, in

which time was the repeated measure along the cycle factor, diagnosis was the grouping factor, and neurosteroid concentrations were the dependent variables. This test was done in patients and controls for the three neurosteroid measurements of the drug-free month and then for the three neurosteroid measurements of the paroxetine therapy month in patients. Moreover, an ANOVA for repeated measures was done only in patients to compare the three neurosteroid measurements before and during therapy. A post-hoc Student’s ttest for unpaired data assessed the significance of possible differences between patients and controls of each neurosteroid at each point of the 2 months, and a Student’s t-test for paired data assessed the significance of a possible difference in patients of neurosteroids in the drug-free and in the therapy months. The correlations between psychopathological and neurosteroid data were analyzed using Spearman’s rank order of correlation. Given the exploratory nature of our study, correlations were considered significant for P values-0.05. 3. Results Table 1 presents the mean scores for the PASS, FQtot and STAI scales. Values of PASS scores obtained during the mid-luteal and premenstrual phases decreased significantly after 1 month of therapy (zs3.32, Ps0.0009, zs3.32, Ps0.0002) and those of FQtot also decreased significantly during the first drug-free month from the early follicular to the mid-luteal periods and to the premenstrual phases (days 7 to 22 to 27) (zs 3.55, Ps0.0004, zs3.17, Ps0.001), and during the therapy month they were lower in the earlyfollicular and pre-menstrual phases than those in

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

111

Table 2 Neurosteroid concentrations Neurosteroids (hgyml)

A1

A2

A3

B1

B2

B3

PREG p c

5.8"3.4 4.6"1.7

7.6"3.1 7.9"4.0

5.8"2.4** 3.6"1.3

6.2"4.0

7.9"3.3

5.8"2.3

PROG p c

1.6"0.8 1.4"0.3

9.2"4.8** 5.3"3.6

2.9"2.5 3.5"1.9

2.4"2.3

8.7"4.3**

3.2"2.6

3a,5a-THPROG p c

4.0"1.7* 2.6"0.8

7.3"3.3 7.3"3.8

8.2"3.5 5.7"0.9

3.8"2.3

7.4"2.8

8.9"2.1***

DHEA p c

4.6"2.2 4.1"1.7

4.7"2.7 4.5"2.4

5.6"2.8 4.3"1.4

4.9"3.5

3.9"2.0

4.9"1.6

3a,5a-THDOC p c

2.1"1.1 2.0"0.8

3.0"1.1 3.0"1.2

3.1"1.5** 1.7"0.8

2.1"0.8

2.9"1.3

3.2"1.2**

Means"S.D.; p, patients; c, controls; PROG, progesterone; PREG, pregnenolone; 3a,5a-THPROG, allopregnanolone; DHEA, dehydroepiandrosterone; 3a,5a-THDOC, tetrahydrodeoxycorticosterone; A1 , drug-free cycle, day 7; A2 , drug-free cycle, day 22; A3, drug-free cycle, day 27; B1, therapy cycle; day 7; B2, therapy cycle, day 22; B3 , therapy cycle; day 27; Patients vs. controls, *P-0.01; **P-0.02; ***P-0.002.

the drug-free month (zs2.66, Ps0.007; zs3.23, Ps0.001). STAI values in the therapy month decreased from the follicular to the mid-luteal to the premenstrual phases (zs3.23, Ps0.001), which were also significantly lower than in the drug-free month (zs2.41, Ps0.01; zs2.66, Ps 0.007). Table 2 presents the mean values for the neurosteroid concentrations. ANOVA for repeated measures for PREG values between patients and controls revealed a significant effect of time in the drug-free month (Fs9.66, Ps0.01) and in the paroxetine therapy month (Fs14.4, Ps0.0001) but not of treatment and treatment per time in both months. The post-hoc Student’s t-test for unpaired data showed that the difference between patients and controls was significant only at the premenstrual phase of the drug-free month (ts2.41, d.f.s 22, Ps0.02). When values pretherapy were analyzed against those after therapy in patients, ANOVA for repeated measures revealed a significant effect of time (Fs6.08, Ps0.005), but not of treatment and treatment per time. Student’s t-test for paired data revealed no significant differences

in the concentrations of the neurosteroid before and after therapy. ANOVA for repeated measures for PROG values during the drug-free month between patients and controls revealed a significant effect of time (Fs 21.32, Ps0.01), of diagnosis (Fs6.56, Ps0.02) and of diagnosis per time (Fs4.26, Ps0.03). Student’s t-test analysis for unpaired data showed that the difference between patients and controls was significant only at the luteal phase of the cycle (ts2.43, d.f.s34, Ps0.02). During the month of paroxetine treatment, ANOVA revealed only a significant effect of time (Fs14.4, Ps 0.0001), but not of diagnosis and diagnosis per time. The Student’s t-test analysis showed that the difference between patients and controls was significant at the mid-luteal phase of the cycle (ts 2.3, d.f.s34, Ps0.02). When data after therapy were analyzed against those before therapy, ANOVA revealed a significant effect of time (Fs28.5, Ps0.01) but not of treatment and treatment per time. Student’s t-test showed that there were no differences between the neurosteroid concentrations at each point of the 2 months.

112

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

ANOVA for repeated measures for 3a,5aTHPROG values in the drug-free month of patients and controls revealed a significant effect of time (Fs14.83, Ps0.01) and of diagnosis (Fs5.06, Ps0.04) but not of diagnosis per time. Student’s t-test revealed that the difference between patients and controls was significant at the follicular phase (ts2.49, d.f.s34, Ps0.01). During the therapy month, ANOVA revealed a significant effect of time (Fs20.3, Ps0.001), but not of treatment and of treatment per time. Student’s t-test revealed that the difference between patients and controls was significant only at the premenstrual phase (ts 3.5, d.f.s20, Ps0.002). The ANOVA between values during the drug-free and the paroxetine month revealed a significant effect of time (Fs 28.5, Ps0.01), but not of treatment and treatment per time. Student’s t-test showed that there were no significant differences at any time of the 2 months. ANOVA for repeated measures for DHEA values between patients and controls revealed no significant effect of time, diagnosis or diagnosis per time during the drug-free and the paroxetine months. Pretherapy values and values obtained during therapy in patients also revealed no significant effect of time, of treatment or of treatment per time. The same was shown with Student’s ttest. ANOVA for repeated measures for 3a,5aTHDOC in patients and controls revealed that during the drug-free month there was a significant effect of time (Fs4.81, Ps0.02) but not of diagnosis and diagnosis per time, while during the paroxetine therapy month there was a significant effect of time (Fs4.9, Ps0.01) and of treatment per time (Fs5.18, Ps0.01) but not of treatment. Student’s t-test analysis showed that the difference between the two groups was significant only at the premenstrual phase of both months (drug-free month: ts2.34, d.f.s22, Ps0.02; paroxetine month: ts2.45, d.f.sPs0.02). ANOVA analysis of data before and during therapy revealed that there was a significant difference for time (Fs 4.8, Ps0.02) but not for treatment and treatment per time. Student’s t-test revealed no significant differences at any point of the 2 months.

Spearman’s rank order correlation coefficients revealed that DHEA concentrations correlated positively with PASS scores at the follicular phase of the drug-free month (A1: rs0.4, P-0.04), positively with the FQtot scores at the follicular phases of the drug-free and of the paroxetine month (A1 and B1: rs0.38, P-0.05) and negatively with the STAI scores at the premenstrual phase of the paroxetine month (B3 rsy0.70, P-0.015). Concentrations of 3a,5a-THDOC correlated positively with the FQtot scores at the follicular and at the mid-luteal phases of the drug-free (A1 and A2: rs 0.42, P-0.03). The above-mentioned correlations were no longer significant after Bonferroni correction. 4. Discussion and conclusions The major findings of our study regard both the symptomatological and the biochemical aspects of PD. During the drug-free month of observation in patients, the severity of phobic symptomatology, measured by the FQtot scores, was significantly lower in the mid-luteal and premenstrual phases than in the early-follicular phase of the cycle, while PASS and STAI scores, which measure panic symptomatology and level of anxiety, respectively, were not significantly different in the three phases. Given that patients had not been treated with benzodiazepines for at least 25 days and SSRI therapy had been suspended for at least 6 months before the study began, the significant improvement of the phobic symptomatology observed in the mid-luteal and premenstrual phases of the drugfree month could not be ascribed to carry-over effects of previous pharmacological treatment. In the following month of therapy, PASS scores were significantly lower in the premenstrual phase than in the early follicular and mid-luteal phases. Since the patients were being treated with paroxetine, however, it is difficult to discriminate between the pharmacologically induced improvement and one possibly due to the cycle-related changes of gonadal steroids. The spontaneous symptomatological improvement in the mid-luteal phase of the drugfree month is in line with our previous finding of a reduction of CO2-induced anxiety during the luteal phase of the cycle (Perna et al., 1995).

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

These data suggest that a specific biochemical mechanism might play a role in the post-ovulatory phase of the cycle to buffer the severity of panic symptomatology. This might be related to the surge in the luteal phase of PROG and its GABAergic metabolites, whose anxiolytic effects have been repeatedly observed in animal models of anxiety (Wieland et al., 1991; Kavaliers et al., 1994; Bitran et al., 1995, 2000; Picazo and Fernandez-Guasti, 1995; Brot et al., 1997; Barbaccia et al., 2000). Our results also show that the pattern of secretion of the neurosteroids is altered, the 3a,5aTHPROG levels being higher than normal in the follicular phase of the drug-free month and in the premenstrual phase of the therapy month, PROG levels being higher than normal in the mid-luteal phases of the drug-free and therapy months, PREG levels being higher than normal in the premenstrual phases of the drug-free and therapy months, and 3a,5a-THDOC in the premenstrual phase of the drug-free month. In contrast, DHEA secretion in each of the three phases of both months was not different from that of controls. The higher than normal concentrations of neurosteroids in the panic patients might be the expression of compensatory mechanisms involving stress reactions. Even though the stress disorder hypothesis of PD is controversial, since spontaneous and pharmacologically induced panic attacks have been reported to be accompanied by activation of the hypothalamo-pituitary-adrenal (HPA) axis by some researchers (Koszycki et al., 1998; Bandelow et al., 2000) but not by others (Cameron et al., 1987; Sinha et al., 1998; Geraci et al., 2002), hyperactivity of HPA function in the interictal phases of PD suggesting the existence of an underlying activation of the axis has been reported by some investigators (Brambilla et al., 1991; Abelson and Curtis, 1996; Schreiber et al., 1996). It has been shown that a variety of stressors induces surges of PROG and of other neurosteroids in rat brain and plasma and in human plasma (Frye et al., 1996; Purdy et al., 1991b; Elman and Breier, 1997; Barbaccia et al., 2001), an effect due to the activation of the HPA axis with corticotropinreleasing hormone (CRH) stimulating neurosteroid surges from the adrenals (Elman and Breier, 1997). PROG and its metabolites, in turn, may act as

113

endogenous stress-protective compounds, since they inhibit the synthesis and release of CRH and the activation of the HPA axis (Patchev et al., 1994). This feedback mechanism might counteract the well-known anxiogenic activity of CRH (Holsboer, 1989; Dunn and Berridge, 1990; Contarino and Gold, 2002) and interrupt a potentially dangerous prolonged activation of the system. The fact, however, that the neurosteroid hypersecretion occurs only in some and not all phases of the cycles suggests that it might be ovary- and not adrenal-dependent, which seems to exclude that their surge might be stress-related. Similarly, we failed to find changes in DHEA concentrations in our patients in each phase of the drug-free and therapy months. In fact, the reported HPA hyperfunction should have resulted in adrenal-dependent hypersecretion of this steroid. We can suggest that the high concentrations of 3a,5a-THPROG in the early follicular phase, of PROG in the mid-luteal phase, and of PREG, 3a,5a-THPROG and 3a,5aTHDOC in the premenstrual phase could have inhibited any eventual DHEA hypersecretion. The observation that neurosteroid alterations were dissociated one from the other in the three phases of the menstrual cycles is intriguing. Enzymatic alteration involving the non-P450-linked 5areductase and 3a-hydroxysteroid-dehydrogenase enzymatic activities, or impaired clearance of the neurosteroids, might be responsible for this phenomenon, even though at the moment we have no proof of it. Our preliminary observations show that DHEA and 3a,5a-THDOC concentrations correlate with anxiety, panic and fear. Even though the Bonferroni analysis excludes the significance of these correlations, due to the large number of analyses done on biological and psychological parameters, we decided to report these data at least as a preliminary observation, which might become statistically significant by investigating a larger population of patients. What is intriguing is that the correlations we have observed between biological and psychopathological parameters do not occur at each point of the two cycles. This is intriguing, even though suggesting a relation between the anxiety disorder and the secretion of neurosteroids leaves open the question of the real significance

114

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

of the biochemical impairments in the context of the pathogenesis of PD and its symptomatology. Possibly, a larger group of patients and a longer observation could clarify this matter. The lack of effects of paroxetine administration on neurosteroid secretion is also intriguing, since SSRIs have been demonstrated to increase previously lower than normal neurosteroid secretion in patients with MDD (Strohle et al., 2002). The phenomenon was already observed by Strohle et al. (2002), who suggested that the ineffectiveness of paroxetine treatment might be attributed to the already higher than normal baseline concentrations of neurosteroids. Given that we measured neurosteroid concentrations only once in the early follicular, mid-luteal and premenstrual phases of both the drug-free and therapy months and in controls during 1 month, and that the sample examined was relatively small, the validity of our data needs to be confirmed. In conclusion, we suggest that in PD the stressful severe anxiety responsible for an underlying chronic state of HPA hyperactivity is followed by a neurosteroid hypersecretion, which might develop in the attempt to inhibit the HPA hyperfunction and the CRH-linked severe anxiety. Since the higher than normal levels of the neurosteroids are not able to block the PD symptomatology, other impairments of this buffering system might be present, possibly including altered receptor sensitivity andyor second messenger function. References Abelson, J.L., Curtis, J.C., 1996. Hypothalamo-pituitary-adrenal axis activity in panic disorder. Archives of General Psychiatry 53, 323–331. Argyle, N., Deltito, J., Allerup, P., Albus, M., Nutziger, D., Rasmussen, S., Ayuso, J.L., Bech, P., 1991. The panicassociated symptoms scale: measuring the severity of panic disorder. Acta Psychiatrica Scandinavica 83, 20–31. Bandelow, B., Wedeking, D., Pauls, J., Brooks, A., Hajak, G., Ruther, E., 2000. Salivary cortisol in panic attaks. American Journal of Psychiatry 157, 454–456. Barbaccia, M.L., Lello, S., Sidiropoulou, T., Cocco, T., Sorge, S.P., Cocchiarale, A., Piermarini, V., Sabato, A.F., Trabucchi, M., Romanini, C., 2000. Plasma 5a-androstane-3a-,17bdiol, an endogenous steroid that positively modulates GABA receptor function and anxiety: a study in menopausal women. Psychoneuroendocrinology 25, 659–675.

Barbaccia, M.L., Serra, M., Purdy, R.H., Biggio, G., 2001. Stress and neuroactive steroids. In: Biggio, G., Purdy, R.H. (Eds.), Neurosteroids and Brain Function. International Review of Neurobiology, vol. 46. Academic Press, San Diego, pp. 441–477. Bernardi, F., Lanzone, A., Cento, R.M., Spada, R.S., Pezzani, I., Genazzani, A.D., Luisi, S., Luisi, M., Petraglia, F., Genazzani, A.R., 2000. Allopregnanolone and dehydroepiandrosterone response to corticotropin-releasing factor in patients suffering from Alzheimer’s disease and vascular dementia. European Journal of Endocrinology 142, 466–471. Bicikova, M., Dibbelt, L., Hill, M., Hampl, R., Starka, L., 1998. Allopregnanolone in women with premenstrual syndrome. Hormone and Metabolism Research 30, 227–230. Biggio, G., Purdy, R.H., 2001. Neurosteroid and Brain Function. International Review of Neurobiology, vol. 46. Academic Press, San Diego. Bitran, D., Klibansky, D.A., Martin, G.A., 2000. The neurosteroid pregnanolone prevents the anxiogenic-like effect of inescapable shock in the rat. Psychopharmacology 151, 31–37. Bitran, D., Shiekh, M., McLeod, M., 1995. Anxiolytic effect of progesterone is mediated by the neurosteroid allopregnanolone at brain GABAA receptors. Journal of Neuroendocrinology 7, 171–177. Brambilla, F., Bellodi, L., Perna, G., Battaglia, M., Sciuto, G., Diafferia, G., Petraglia, F., Panerai, A., Sacerdote, P., 1991. Psychoimmunoendocrine aspects of panic disorder. Biological Psychiatry 2, 581–583. Breier, A., Charney, D., Heninger, G., 1986. Agoraphobia with panic attacks: development, diagnostic stability, and course of illness. Archives of General Psychiatry 43, 1029–1036. Brot, M.D., Akwa, Y., Purdy, R.H., Koob, G.F., Britton, K.T., 1997. The anxiolytic-like effects of the neurosteroid allopregnanolone: interactions with GABA-A receptors. European Journal of Pharmacology 325, 1–7. Cameron, O.G., Kuttesch, D., McPhee, K., Curtis, G.C., 1988. Menstrual fluctuation in the symptoms of panic anxiety. Journal of Affective Disorders 15, 169–174. Cameron, O.G., Lee, M.A., Curtis, G.C., McCann, D.S., 1987. Endocrine and physiological changes during spontaneous panic attacks. Psychoneuroendocrinology 12, 321–331. Contarino, A., Gold, L.H., 2002. Targeted mutations of the corticotropin-releasing factor system: effects on physiology and behavior. Neuropeptides 36, 103–116. Cook, B.L., Noyes, R., Garvey, M.J., Beach, V., Sobotka, J., Chaudhry, D., 1990. Anxiety and the menstrual cycle in panic disorder. Journal of Affective Disorders 19, 221–226. Cowley, D.S., Roy-Byrne, P.P., 1989. Panic disorder during pregnancy. Journal of Psychosomatic Obstetrics and Gynecology 10, 193–210. Crawley, J.N., Glowa, J.R., Majewska, M.D., Paul, S.M., 1986. Anxiolitic activity of an endogenous adrenal steroid. Brain Research 398, 382–385. Daly, R.C., Kim, M.Y., Bloch, M., Schmidt, P.J., Purdy, R.M., Rubinow, D.R., 2000. Neurosteroids in reproductive endocrine-related mood disorders (Abstract). Biological Psychiatry 47, 5S.

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116 Deci, P.A., Lydiard, R.B., Santos, A.B., Arana, G.W., 1992. Oral contraceptives and panic disorder. Journal of Clinical Psychiatry 53, 163–165. Dunn, A.J., Berridge, C.W., 1990. Physiological and behavioral responses to corticotropin-releasing factor administration: is CRF a mediator of anxiety or stress responses? Brain Research Reviews 15, 71–100. Elman, I., Breier, A., 1997. Effects of acute metabolic stress on plasma progesterone and testosterone in male subjects: relationship to pituitary–adrenocortical axis activation. Life Sciences 61, 1705–1712. Flood, J.F., Baker, M.L., Davis, J.L., 1990. Modulation of memory processing by glutamic receptor agonists and antagonists. Brain Research 521, 197–202. Flood, J.F., Morley, J.E., Roberts, E., 1992. Memory-enhancing effects in male mice of pregnenolone and steroids metabolically derived from it. Proceedings of the National Academy of Sciences of the United States of America 89, 1567–1571. Flood, J.F., Smith, G.E., Roberts, E., 1988. Dehydroepiandrosterone and its sulfate enhance memory retention in mice. Brain Research 447, 269–278. Frye, C.A., McCormik, C.M., Coopersmith, C., Erskine, M.S., 1996. Effects of paced and non-paced mating stimulation on plasma progesterone, 3a-diol and corticosterone. Psychoneuroendocrinology 21, 431–439. George, D.T., Ladenheim, J.A., Nutt, D.J., 1987. Effect of pregnancy on panic attacks. American Journal of Psychiatry 144, 1078–1079. George, M.S., Guidotti, A., Rubinow, D., Pan, B., Mikalauskas, K., Post, R.M., 1994. CSF neuroactive steroids in affective disorders: pregnenolone, progesterone and DBI. Biological Psychiatry 35, 775–780. Geraci, M., Anderson, T.S., Slate-Cothren, S., Post, R.M., McCann, U.D., 2002. Pentagastrin-induced sleep panic attacks: panic in the absence of elevated arousal. Biological Psychiatry 52, 1183–1189. Heuser, I., Deuschle, M., Luppa, P., Schweiger, U., Standhardt, H., Weber, H., 1998. Increased diurnal plasma concentrations of dehydroepiandrosterone in depressed patients. Journal of Clinical Endocrinology and Metabolism 83, 3130–3133. ´ Heydari, B., Le Melledo, J.M., 2002. Low pregnanolone sulphate plasma concentrations in patients with generalized social phobia. Psychological Medicine 32, 929–933. Holsboer, F., 1989. Psychiatric implications of altered limbichypothalamic-pituitary-adrenocortical activity. European Archives of Psychiatry and Neurological Science 238, 302–322. Kavaliers, M., Wiebe, J.P., Galea, L.A., 1994. Reduction of predator odor-induced anxiety in mice by the alpha-hydroxy4-pregnen-20-one (3 alpha HP). Brain Research 9, 325–329. Klein, D.F., 1964. Delineation of two drug-responsive anxiety syndromes. Psychopharmacologia 5, 397–408. Klein, D.F., 1993. False suffocation alarms, spontaneous panics, and related conditions. An integrative hypothesis archives of General Pyschiatry 50, 306–317.

115

Klein, D.F., Skrobala, A., Garfinkel, R.S., 1995. A preliminary look at the effects of pregnancy on the course of panic disorder. Anxiety 1, 227–232. ´ Koszycki, D., Zacharko, R.M., Le Melledo, J.M., Bradwejn, J., 1998. Behavioral, cardiovascular, and neuroendocrine profiles following CCK-4 challenge in healthy volunteers: a comparison of panickers and nonpanickers. Depression and Anxiety: Anxiety 8, 1–7. Marks, I.M., Mathews, A.M., 1979. Brief standard self-rating for phobic patients. Behaviour Research and Therapy 17, 263–267. Metz, A., Sichel, D.A., Goff, D.C., 1988. Postpartum panic disorder. Journal of Clinical Psychiatry 49, 278–279. Monteleone, P., Tortorella, A., Luisi, M., Maj, M., Genazzani, A.R., 2001. Peripheral neurosteroid concentrations in eating disorders. Psychosomatic Medicine 63, 62–68. Patchev, V.K., Shoaib, M., Holsboer, F., Almeida, O.F.X., 1994. The neurosteroid tetrahydroprogesterone counteracts corticotropin-releasing hormone-induced anxiety and alters the release and gene expression of corticotropin-releasing hormone in the rat hypothalamus. Neuroscience 62, 265–271. Perna, G., Brambilla, F., Arancio, C., Bellodi, L., 1995. Menstrual cycle-related sensitivity to 35% CO2 in panic patients. Biological Psychiatry 37, 528–532. Picazo, O., Fernandez-Guasti, A., 1995. Anti-anxiety effects of progesterone and some of its reduced metabolites: an evaluation using the burying behavior test. Brain Research 680, 135–141. Price, W.A., Heil, D., 1988. Estrogen-induced panic attacks. Psychosomatics 29, 433–435. Purdy, R.H., Morrow, A.L., Blinn, J.R., Paul, S.M., 1991. Synthesis, metabolism and pharmacological activity of 3ahydroxysteroid which potentiate GABA-receptor-mediated chloride ion uptake in rat cerebral cortical synaptoneurosomes. Journal of Medical Chemistry 33, 1572–1581. Purdy, R.H., Morrow, A.L., Moore Jr., P., Paul, S.M., 1991. Stress-induced elevation of a-amino-butyric acid type A receptor-active steroid which potentiate GABA-receptormediated chloride ion uptake in rat cerebral cortical synaptoneurosomes. Journal of Medical Chemistry 33, 1572–1581. Rapkin, A.J., Morgan, M., Goldman, L., Brann, D.V., Simone, D., Mahesh, V.B., 1997. Progesterone metabolite allopregnanolone in women with premenstrual syndrome. Obstetrics and Gynecology 90, 709–714. Rasgon, N., Serra, M., Biggio, G., Pisu, M.G., Fairbanks, L., Tanavoli, S., Rapkin, A., 2001. Neuroactive steroid-serotonergic interaction: responses to an intravenous L-tryptophan challenge in women with premenstrual syndrome. European Journal of Endocrinology 145, 25–33. Romeo, E., Strohle, A., Spalletta, G., Di Michele, F., Herman, B., Holsboer, F., Pasini, A., Rupprecht, R., 1998. Effects of antidepressant treatment on neuroactive steroids in major depression. American Journal of Psychiatry 155, 910–913.

116

F. Brambilla et al. / Psychiatry Research 118 (2003) 107–116

Rupprecht, R., Holsboer, F., 2001. Neuroactive steroids in neuropsychopharmacology. In: Biggio, G., Purdy, R.H. (Eds.), Neurosteroids and Brain Function. International Review of Neurobiology, vol. 46. Academic Press, San Diego, pp. 441–477. Sandberg, D., Fyer, A.J., Endicott, J., 1986. Premenstrual changes in anxiety patients. (Abstract) Proceedings of the Annual Meeting of the American Psychiaric Association. Schreiber, W., Lauer, C.J., Krumrey, K., Holsboer, F., Krieg, J.C., 1996. Dysregulation of the hypothalamic-pituitaryadrenocortical system in panic disorder. Neuropsychopharmacology 15, 7–15. ´ Semeniuk, L., LeMelledo, J.M., Shangri, G.S., 2001. Neuroactive steroid levels in patients with generalized anxiety disorder. Journal of Neuropsychiatry and Clinical Neuroscience 13, 396–398. Serra, M., Madau, M.F., Chessa, M., Caddeo, M., Sanna, E., Trapani, G., Franco, M., Liso, G., Purdy, R.H., Barbaccia, M.L., Biggio, G., 1999. 2-Phenyl-imidazol-w1,2-ax-pyridine derivatives as ligands for peripheral benzodiazepine receptors: stimulation of neurosteroid synthesis and anticonflict action. British Journal of Pharmacology 127, 177–182. Sheehan, D.V., Lecrubier, Y., Janavs, J., Knapp, E., Weiler, E., Bonora, P., Sheedan, M., Amorim, P., Baker, R., Sheehan, K.H., Lepine, J.P., 1994. Mini International Neuropsychiatric Interview, ParisyTampa. Sinha, S.P., Coplan, J.D., Pine, D.S., Martinez, J.A., Klein, D.F., Gorman, J.M., 1998. Panic induced by carbon dioxide inhalation and lack of hypothalamic-pituitary-adrenal axis activation. Psychiatry Research 86, 93–98. Spielberger, C.D., Gorsuch, R., Lushere, R., 1970. STAI. Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press, Palo Alto, CA. Spivak, B., Maayan, R., Kotler, M., Mester, R., Gil-Ad, R., Shtaif, B., Wizman, A., 2000. Elevated circulatory level of GABA(A)-antagonistic neurosteroids in patients with combat-related post-traumatic stress disorder. Psychological Medicine 30, 1227–1231. Stein, M.B., Schmidt, P.J., Rubinow, D.R., Uhde, T.W., 1989. Panic disorder and the menstrual cycle: panic disorder patients, healthy control subjects, and patients with premenstrual syndrome. Archives of General Psychiatry 146, 1299–1303. Strohle, A., Romeo, E., Di Michele, F., Pasini, A., Yassouridis, A., Holsboer, F., Rupprecht, R., 2002. GABA-A receptormodulating neuroactive steroid composition in patients with panic disorder before and during paroxetine treatment. American Journal of Psychiatry 159, 145–148.

Strohle, A., Romeo, E., Hermann, B., Pasini, A., Spalletta, G., Di Michele, F., Holsboer, F., Rupprecht, R., 1999. Concentrations of 3a-reduced neuroactive steroids and their precursors in plasma of patients with major depression and after clinical recovery. Biological Psychiatry 45, 274–277. Strohle, A., Romeo, E., Holsboer, F., Rupprecht, R., 2000. Neuroactive steroids in major depression and after antidepressant treatment. Neuropsychopharmacology 23, S26 Abstract. Tait, G.R., McManus, K., Bellavance, F., Lara, N., Chrapko, ´ W., Le Melledo, J.M., 2002. Neuractive steroid changes in response to challenge with the panicogenic agent pentagastrin. Psychoneuroendocrinology 27, 417–429. Uzunova, V., Sheline, Y., Davis, J.M., Rasmusson, A., Uzunov, D.P., Costa, E., Guidotti, A., 1998. Increase in the cerebrospinal fluid content of neurosteroids in patients with unipolar major depression who are receiving fluoxetine or fluvoxamine. Proceedings of the National Academy of Sciences of the United States of America 95, 3239–3244. Villeponteaux, V.A., Lydiard, R.B., Laraia, M.T., Stuart, G.W., Ballenger, J.C., 1992. The effects of pregnancy on preexisting panic disorder. Journal of Clinical Psychiatry 53, 201–203. Wang, M.D., Backstrom, T., Sundstrom, T., Wahlstrom, G., Olsson, T., Zhu, D., Johansson, I.M., Bjorn, I., Bixo, M., 2001. Neuroactive steroids and central nervous system disorders. In: Biggio, G., Purdy, R.H. (Eds.), Neurosteroid and Brain Function. International Review of Neurobiology, vol. 46. Academic Press, San Diego, pp. 421–459. Wang, M.D., Seippel, L., Purdy, R.H., Backstrom, T., 1996. Relationship between symptom severity and steroid variation in women with premenstrual syndrome. Study on serum pregnenolone, pregnenolone sulfate, 5a-pregnan-3,20-dione and 3a-hydroxy-5a-pregna-20-one. Journal of Clinical Endocrinology and Metabolism 81, 1076–1082. Wieland, S., Lan, N.C., Mirasedeghi, S., Gee, K.W., 1991. Anxiolytic activity of the progesterone metabolite 5a-pregnan-3a-ol-20-one. Brain Research 565, 263–268. Wolkowitz, O.M., Reus, V.I., Keebler, A., Nelson, N., Friedland, M., Brizendine, L., Roberts, E., 1999. Double-blind treatment of major depression with dehydroepiandrosterone. American Journal of Psychiatry 156, 646–649. Wolkowitz, O.M., Reus, V.I., Roberts, E., Manfredi, F., Chan, T., Raum, V.J., Ormiston, S., Johnson, R., Canick, J., Brizendine, L., Weingartner, H., 1997. Dehydroepiandrosterone (DHEA) treatment of depression. Biological Psychiatry 41, 311–318. Young, A.H., Gallagher, P.B., Porter, R.J., 2002. Elevation of the cortisolydehydroepiandrosterone ratio in drug-free depressed patients. American Journal of Psychiatry 159, 1237–1239.